Heat-activation of TRPM5 underlies thermal sensitivity of sweet taste. Supplementary information.

Heat activation of TRPM5 is membrane-delimited.

We tested whether heat activation of TRPM5 was present in cell-free inside-out patches. In most of the cases, TRPM5 currents rapidly decay to levels indistinguishable from the background currents after patch excision. However, in a very little number of cases, there was some TRPM5 current left to be analysed for temperature dependence (Fig. S1).

Supplementary Figure S1. Heat activation of TRPM5 in inside-out patches.

(a) TRPM5 currents (right panel) elicited by a voltage ramp from –75 to +150 mV, with a holding potential of –25 mV (left panel). The extracellular solution contained (in mM): 50 NaCl, 100 NMDG+, 5 CaCl2, 1 MgCl2 and 10 HEPES and titrated to 7.4 with NaOH. The intracellular solutions contained 150 mM NaCl, 10 mM HEPES, titrated to pH 7.2 with NaOH.

(b) Temperature of the bath perfusion during the course of the experiment.

(c) Time course of amplitude of currents at –60 and +150 mV. After patch excision in the presence of 100 µM Ca2+ in the bath (intracellular) solution (see arrow), TRPM5 currents were activated rapidly and then decayed to a steady-state value (dashed line) within 1 min. Cooling reversibly decreased both inward and outward currents. Horizontal bars denote periods of perfusion without intracellular Ca2+ (HEDTA 2 mM), during which TRPM5 activation was abolished.

(d) Temperature dependence of current amplitude at –60 and +150 mV. Similar results were obtained in three other cells (data not shown).

Temperature dependence of TRPM4

The modulation of TRPM4 gating by temperature was qualitatively similar to that in TRPM5. Increasing temperatures strongly shifted the activation curve in the negative direction along the voltage axis and increased the rate of current relaxation at every membrane potential (Fig. S2a, c-e). The Ca2+ concentration for half-maximal activation (EC50) decreased upon heating with an estimated Q10 of 1.36 ± 0.02 (Fig. S2f).

Supplementary figure S2. Temperature dependence of TRPM4.

(a) Typical current traces evoked by depolarizing pulses from -25 to +175 mV from a holding potential of -50 mV at three different temperatures.

(b) Current traces at different temperatures predicted by the model for TRPM4.

(c-d) Average activation curves and voltage dependence of the time constant of current relaxation at different temperatures (n = 4-7).

(f) Vact as a function of temperature.

Continuous lines in (d and e) correspond to the fit of the two-state gating model and in (c) to the activation curves predicted by the model.

(f) Dose responses to intracellular Ca2+ at different temperatures (n = 3-5). Continuous lines are the fit of the experimental data with Hill functions of the form: , where EC50 is the effective concentration for 50% current activation and N is the Hill coefficient. The Q10 of EC50 was determined by fitting temperature dependence of EC50 with the equation: , where Tmax = 37 °C and EC50(Tmax) = 107 µM.. The inset shows the temperature dependence of the effective concentration for 50% activation with intracellular Ca2+.

Modelling the gating of TRPM4 and TRPM5.

In the presence of saturating intracellular Ca2+ concentrations, the gating of TRPM4 and TRPM5 can be described in terms of a two-states (closed-open) model15. Within this approximation, rate constants for channel opening (α) and closing (β) are only voltage- and temperature-dependent and can be described according to Eyring’s rate theory11:

and

where k is the Boltzmann constant, T is the absolute temperature, h is Planck’s constant, ΔHi and ΔSi (i = α, β) are the enthalpy and entropy (per mole) changes associated with each transition, z is the effective gating valence of the voltage sensor, R is the gas constant, F is Faraday’s constant and  accounts for the coupling between the local electric potential sensed by the gating charge and the membrane potential V. This model predicts that voltage for half-maximal activation (Vact) and the slope factor of the activation (sact) are given by: and , where ΔH‡ = ΔHα – ΔHβ and ΔS‡ = ΔSα – ΔSβ. Fitting these expressions to the temperature dependence of Vact (Figs. 2f and S2e) and sact (data not shown) we obtained ΔH‡, ΔS‡and z. Then, we used these values to fit simultaneously the voltage dependences of the time constant of current relaxation at different temperatures (Figs. 2e and S2d). By writing this equation in the following form:

,

we were able to obtain ΔHα, ΔSα, and . Then, values of ΔHβand ΔSβ where calculated by ΔHβ = ΔHα– ΔH‡and ΔSβ= ΔSα– ΔS‡.

We observed some discrepancies between the experimental data and the model predictions at high temperatures (Fig. 2e and S2d). This may reflect the fact that the two-state model is indeed the simplest approximation to the gating properties of TRPM4 and TRPM5.

For both TRPM4 and TRPM5, the rate constant of channel opening (activation, α) is more temperature dependent than that of channel closing (deactivation, β) (Fig. S3). This corresponds to what we have previously shown for heat activated TRPV13.

Supplementary Figure S3.Predictions of the kinetic models for TRPM4 and TRPM5.

Arrhenius plots of the rate constants of opening (α) and closing (β) at different membrane potentials obtained from the models for TRPM4 (a) and TRPM5 (b).